To demonstrate proof of concept for this strategy, we have initiated studies to identify immunogenic T cell epitopes for mutated tumor antigens, such as K-ras which expressed by several GI cancers. A library of overlapping 9mer and 10mer peptides which span the amino acid mutation has been synthesized and we are currently performing T cell stimulations with PBL of a variety of HLA types to determine the immunogenicity of the peptides and their MHC restriction in patients whose tumors possess k-ras mutations. In contrast to low throughput screening which require several rounds of in vitro sensitization to evaluate each peptide, our current approach will screen hundreds of individual T cell cultures one week after peptide stimulation. A significant aspect of this project is to apply automated instrumentation for high throughput T cell stimulation and screening in a 384 well format. Further refinements include the comparative evaluation of computer based algorithms that predict the immunogenicity and potential proteosomal cleavage sites for the putative epitopes. This in silico screening would complement our high throughput functional screening of epitopes.